2007 OPEN FORUM Abstracts
LITERATURE SEARCH RESULTS: PUBMED AND GOOGLE SCHOLAR IN COMPARISON TO COCHRANE EVIDENCE-BASED SYSTEMATIC REVIEWS
M. E. Anders1, A. Bittle1, B. Free1, T. Jones1, T. Tappan1
Background: Literature searches are essential to evidence-based respiratory care. To conduct literature searches, respiratory therapists rely on search engines, such as PubMed and Google Scholar, to access, filter, and retrieve information. However, the efficiency of these search engines for clinical questions in respiratory care is unknown.
Purpose: The purpose of our study was to compare PubMed and Google Scholar search results for clinical questions in respiratory care to that of a “gold standard”, Cochrane evidence-based systematic reviews.
Methods: We formulated three clinical questions: (a) Does exercise improve outcomes in cystic fibrosis?; (b) Does pulmonary rehabilitation improve outcomes in COPD?; and (c) Does non-invasive positive pressure ventilation improve outcomes in acute exacerbation of COPD?. For each clinical question, a relevant Cochrane Collaboration evidence-based systematic review, which used neither the PubMed nor Google Scholar search engine, served as a “gold standard” for search terms, search limits, and search results, including original research manuscripts and excluding abstracts, conference proceedings, and non-English language publications. We conducted literature searches by applying these search terms and limits, as possible, in PubMed and Google Scholar and compared the search results with that of the corresponding Cochrane evidence-based systematic review. We calculated sensitivity, positive predictive value, specificity, negative predictive value, and positive and negative likelihood ratios.
Results: PubMed demonstrated a higher sensitivity and positive predictive value than Google Scholar for all three literature searches. The table presents these results.
Conclusions: The study results provide evidence to suggest that PubMed is a more efficient search engine than Google Scholar for clinical questions in respiratory care. Compared to Google Scholar, PubMed allows users more control in filtering the search results. PubMed appears to be preferable when respiratory therapists need to conduct efficient yet valid, comprehensive literature searches for evidence-based patient care plans or protocols.
|CF and Exercise||COPD and Rehab.||COPD
|PubMed||5/5 = 100%||23/28 = 82%||7/10 = 70%|
|4/5 = 80%||3/28 = 11%||6/10 = 60%|
|PubMed||5/5 = 100%||18/28 = 64%||7/10 = 70%|
|3/5 = 60%||3/28 = 11%||3/10 = 30%|
|Positive Predictive Value|
|PubMed||5/30 = 17%||23/379 = 6%||7/278 = 3%|
|4/166 = 2%||3/573 = 0.5%||6/291 = 2%|
|Full-text Positive Predictive Value|
|PubMed||5/30 = 17%||18/379 = 5%||7/278 = 3%|
|3/166 = 2%||3/573 = 0.5%||3/291 = 1%|